Deep and even ultra-deep petroleum resources play a gradually increasing and important role with the worldwide continuous advancement of oil and gas exploration and development. In China, the deep carbonate reservoirs in the Tarim Basin are regarded as the key development areas due to their huge reserves. However, due to the unreasonable design of production pressure differential, some production wells suffered from severe borehole collapse and tubing blockage. Therefore, the main purpose of this paper is to optimize a more practical method for predicting the critical production pressure differential. The commonly used analytical methods with different failure criteria for predicting production pressure differential were summarized. Furthermore, their advantages and disadvantages were analyzed. A new numerical model is established based on the finite element theory in order to make the prediction of production pressure differential more accurate. Additionally, both analytical and numerical methods were applied to evaluating the production pressure differential of deep carbonate reservoirs in the Tarim Basin, and the results were discussed compared with field data. In addition, a series of laboratory tests, including porosity and permeability measurements, electron microscope scanning, XRD for mineral analysis, uniaxial and triaxial compressive strength test, etc., were carried out by using the collected carbonate cores from formations deeper than 7000 m to obtain the input parameters of the simulation such as the rock properties. The experimental results showed that the carbonate rocks exhibited a remarkable brittleness and post-peak strain softening. The calculation results revealed that the Mogi-Coulomb criterion is slightly conservative; however, it is more suitable than other criteria to evaluate pressure differential. Furthermore, it has been confirmed by the field data that the finite element numerical method can not only reveal the instability mechanism of the wellbore but also predict the critical production pressure differential accurately. Unfortunately, the on-site operators sometimes require a more convenient way, such as an analytical method, to figure out the pressure differential, even though the evaluation of the numerical method is more accurate. Therefore, the discussion in this paper can provide a basis for the operators to determine the production pressure differential flexibly.